Maximum Entropy Modeling of Indiana Bat (Myotis Sodalis) Maternity Roost Habitat

Publication Year:
2015
Usage 63
Downloads 39
Abstract Views 24
Repository URL:
https://bearworks.missouristate.edu/theses/1343
Author(s):
Lemen, Joseph Robert
Tags:
Indiana bat; Myotis sodalis; landscape; maternity roost; reproduction; MaxEnt; Biology
artifact description
Since 1967, the Indiana bat (Myotis sodalis) has been on the U.S. Endangered Species list due to disruption of hibernating bats in caves, summer habitat degradation, and more recently, the onset of White-nose Syndrome. The purpose of this study is to evaluate landscape variables associated with Indiana bat maternity roost trees in an attempt to better understand what factors play a role in their distribution in north central Missouri. I tracked reproductive female Indiana bats to 20 different primary and secondary roost trees; these are roosts that had multiple bats visit them on multiple occasions. GPS location data for these roosts and 6 environmental parameters (aspect, distance to forest edge, distance to stream or river, elevation, percent tree canopy, and slope) were used as input variables for a MaxEnt model of species distribution. I used ENMTools to identify which analysis features produced the best MaxEnt model for this data set. Linear and quadratic analysis features, separately, fit the data the best. When cross-validated through four replicates, the two models performed equally well with area under the curve (AUC) values of 0.792 and 0.764. Distance to forest edge was the variable with the most influence in both models, followed by elevation and distance to stream. Macro-scale environmental variables provide insight to modeling areas in which Indiana bat maternity roosts might be found in the future. This provides researchers and wildlife managers with a toolset to identify potential habitat to aid in species recovery.